Forum OpenACS Q&A: ACS-JAVA + PostgreSQL

Collapse
Posted by Sean Forbes on
Is there anyone working on an open source solution for the new
acs-java distrobution?  If not, how difficult would it be to port to
PostgreSQL for this system?  Can you estimate how many man hours am I
looking at for such a port.
Collapse
Posted by Roberto Mello on
I'm guessing very few people here have even looked at ACS Java. With ACS Java being licensed under the ADPL, it was a liability for OpenACS developers to even look at its source code.

Red Hat hasn't released details on how it will license ACS Java from now on, so it's all in space so far.

My thinking is that someone would port ACS Java iff $big_bucks were being poured in by some company.

Collapse
Posted by Rocael Hernández Rizzardini on
As aD was adquired by RedHat, I guess that maybe the will want ACS (in this case the Java version) to run under their redhat database (postgres), so, the big bucks will be in the side of Redhat, but no body know certainly that now.

Is there a special reason for what you want to use Java?

Collapse
Posted by Sean Forbes on
I am new to ACS, so i'm not concerned with a legacy ACS solution.  JAVA is the direction that ArsDigita / RedHat seem to be going.  I don't want to start a big JAVA vs. TCL fight.  The truth is that I currently use PHP for web development and that is a lot closer to TCL than JAVA.  JAVA isn't better, but it has some advantges (and disadvantages) over TCL.  The biggest advantage, however, and the reason I am more interested in JAVA, is that ArsDigita / RedHat are moving in that direction.
Collapse
Posted by Don Baccus on
There's no indication whatsoever that RedHat intends to push forward and complete ACS 5.0 Java.  Evidence includes the fact that RedHat has stated their intention to lay off everyone but a handful of folks who work with major accounts such as Siemens (continuing work on their internal Sharenet site, which is ACS Tcl, not Java, based) and the German telephone company.

Unless things have radically changed in the last few weeks, ACS Java is dead as far as RH/aD is concerned.

It's possible that the developers who've been laid off or others might put together an OpenACS-like project to finish it, but I haven't heard  of any such plans.

On the other hand, OpenACS 4.5 Tcl is alive and well and has an active and thriving user and development community.

If you like Java, ACS Java might be a choice though the new platform was never completed AFAIK.  I'd personally recommend seeking a more mainstream Java-based solution if language choice is what's driving your decision.  Forget the ACS and jump into the Java world up to your eyeballs.

On the other hand, if you're attracted to ACS toolkit concepts and language isn't your primary motivation, then there's really no doubt that the OpenACS community is pretty much "it" regarding the ACS these days.

Collapse
Posted by Sean Forbes on
It states on RedHat's Website:

Will Red Hat continue to develop ACS for the Java Platform (ACSJ)?

- Java is an important technology for enterprise customers
-ACSJ provides enterprise customers with a scalable, reliable platform for ACS deployments
- Red Hat has no intention to alter ArsDigita's prior commitment to Java (Note: Sun's trademark prevents us from calling the software "ACS Java", though "ACS for the Java Platform" is OK. Hence the abbreviation "ACSJ".)

Have things changed since this announcement?

Collapse
Posted by Jade Rubick on
I think what Don's getting at is that actions speak louder than
words. Let's say you were buying out WordPerfect, and
continuing development on it. Would you buy the company, and
lay off all the developers? That's rather silly, because you're
getting rid of the very people who know the intricacies of the
code.

I think one thing to look for when you get involved in an open-
source software project is the strength of vitality of the
community behind it. OpenACS does really well in that regard,
and that's one of my main reasons for being involved in it. It's
one of the strongest internet-based groups I've seen.

Collapse
Posted by Don Baccus on
My information comes directly and indirectly from folks involved.

As I said, though, things may've changed recently and RH may be intending to finish ACSJ.  If anyone knows this to be a fact please post!

At the time the acquisition was announced the word was that ACSJ development's dead, other than servicing some existing customers.  I haven't talked to anyone about it in a month or so, though ...

Collapse
Posted by Richard Li on
We have finished the 5.0 release of ACS Java and will be releasing it shortly. It is taking a bit longer because we are all new to Red Hat and there are logistical issues we have to work through (e.g., where to put the software on Red Hat, etc.).
Collapse
Posted by David Walker on
Any news on what license it will be released under?
Collapse
11: aplaws is Java (response to 1)
Posted by Tapiwa Sibanda on
I posted a little while ago about APLAWS, the Accessible and Personalised Local Authority Web Sites project in the UK.

I saw a demo of this at RH/aD offices in London a couple of weeks ago. It does look very slick. A lot of work has gone into the CMS and user interface.

Acording to the products page this will be released sometime later this month.

Although they have said that it will be open sourced, they have made no indication what kind of licence it will be available under. I doubt that it will be too restrictive though, given the UK govt footed the bill via the pathfinder scheme.

APLAWS was designed primarily for use by local authorities, but the data model seems advanced enough to allow use in any CMS/KMS situation. Details are still a bit sketchy, but watch this space.

When I was at their offices, the guys at aD also said that the next stage of the project would be to port it to postgreSQL sometime later this year.

Collapse
Posted by Paul Moody on
I think what Don's getting at is that actions speak louder than words.  Sean, I have been reading the OpenACS and ACS forums like a passion. I started following ACS before there were an OpenACS. I often do not like how Don employs diplomacy. It is not one of his strongest points. But, he most often tells it straight out with no icing. Fortunately for OpenACS, he is most often right. Therefore, actions do speak louder than words.  Consider this in making your decision:

1.    MIT and other universities are strongly supporting OpenACS. See http://dotlrn.mit.edu/

2.    A complete elearning package is being built for OpenACS backed by MIT.

3.    RedHat does support Postgres at a price you may not want to pay. Therefore, I am optimistic that they will finish ACS 5.0 for their postgres release. But, do you think that ACS 5.0 will be for free or will it be priced like their Postgres package?

Now about TCL, I am a man who was and still is just a little sceptical about TCL. In fact, I hate the idea of learning a new scripting or programming language. Although I finally must admit that TCL is not just another scripting language. It is a real programming language with all the benefits of a scripting language. I hate to admit it; it may be better than Java considering all the license problems and the steep learning curve. But TCL is worse than Java if you want heavy support tools like are provide by Borland or SUN. Forgetting support environments, often the only difference in many of these types of programming languages is the hype. TCL has no marketing and no hype, maybe like Don himself. But, once you learn enough about both of them, you have to appreciate the frankness and the functions provided.

The last points, I have looked at ez, uportal, Apache project, OpenCMS, etc. they do not even come close to having the same functions, support or advantages licensing that this weird group and OpenACS can offer. OpenACS could write the book on excellent support and enthusiasm. So again, actions speak louder than words. This group is full of action and honest opinions but not so good on diplomatic words.

Collapse
Posted by Don Baccus on
Releasing APLAWS will be great, sure, but is it the full ACSJ 5.0 toolkit or just customizations on top of CMS and user management and similar core stuff?

That's just a small fraction of the toolkit functionality ... most folks asking whether or not ACSJ 5.0 will be finished are interested in the whole enchilada, not just the CMS package (which is a very important package, of course!)

The first announcement I saw of this also included information about Siemen's ShareNet application.  Because it was coupled with information about ACSJ a reader not familiar with ShareNet would undoubtably assume it had been moved to ACSJ rather than remaining as an ACS Tcl-based site.

In fact it made *me* wonder if it might be true, so I asked around (earlier this week) and was told, no, Siemen's ShareNet is still ACS Tcl.

As far as being diplomatic or not ... I'm just reporting what I've been told, and pointed out that I've not been following things for some time.  Things may've changed.  My information may be obsolete and inaccurate.  Maybe the APLAWS folks are so enthusiastic about ACSJ 5.0 that they've convinced RedHat that finishing it up (the whole toolkit not just CMS and related slices) is a good thing to do.

If so, that would be great.  The more open source toolkits the merrier.  In fact if APLAWS is released under the GPL we may well want to take a close look at the CMS UI, because Ars Digita worked hard to revamp it after getting so many negative comments about the CMS UI in ACS 4 Tcl and I'm sure we could pick up some very good ideas from their effort.

I've not been posting the names of who have told me what because, while no one has asked me not to use their names, no one has told me it is OK to use there names.

Collapse
Posted by Tapiwa Sibanda on
Don, I suspect that the aplaws release is the full enchilada. Here is a snippet from an email I received a couple of days ago from someone within the aplaws mngmt team (in one of the local authorities).
The software will be available within April. We're still doing some enhancements to it. It will be massive - we've thought of maybe making it only available via CD - your thoughts? ....

I am trying to find out more details about packages and licence and will post here as soon as I get something.

Collapse
Posted by John Holroyd on
According to the APLAWS site its all due for release on the 20th April. I hope they dont make it CD-Rom only, that would be quite tedious.
Collapse
Posted by Jun Yamog on
Hi,

I am also interested in ACS Java not because its something that I like but because one of our clients requires us to use it.  I have posted some post here or in aD's bboard.

Here are the facts and opinions that I have so far:

1. ACSJ 5.0 is ACSJ 4.7 with tons of bug fix.  It does not differ from ACSJ 4.7 or 4.6

2. According to one of the ACSJ persistence layer developer it will take around 3 months to port ACSJ to postgres.  porting the create scripts and PDL files.

3. If you are looking for mature OSS CMS solution that is Java based look at mmbase or opencms.  I like mmbase better since it looks more flexible, the community is very active and organized.

4. The permission problem is still present in ACSJ 5.0 so permissions are not used in ACSJ production sites but rather using just the workflow for the permissions.

In my opinion stay away from ACSJ if you don't require it because:

1. According the one of aD's developer it offers no real advantage over OpenACS.  Maybe only bebop is the real advantage.

2. Look at other Java solutions if you need Java.  Why?  ACSJ is not really take advantage of Java platform.  Too much dependence on Oracle not that portable, does not make use of app server scaling.  ACSJ still uses the traditional web farm with one db scaling.  Go look at mmbase and see the difference and advantages of using Java.  mmbase is also a mature solution.

This are just my opinion and the facts that I got.  Some maybe wrong some maybe right.  As for me I may be forced to use ACSJ.

Collapse
Posted by Ming Fai Ma on
redhat has released acs-java 5.0 and it shows it will continue developing the software. and the license is changed to a standard general public license without requirements of stating the product is built on Arsdigita.

So, is there anyone interested in porting acs-java to other dbms?

Collapse
Posted by Talli Somekh on
Ming,

You'll find very little interest here. People are here mostly because of AOLserver and the Tcl base.

talli

Collapse
Posted by Jun Yamog on
Hi Ming,

I am using CCM for one of my projects.  I am working with Red Hat folks... I am also asking this question about a postgres port since I know other people are interested too.

5.1 is running although its internal.  Postgres is being done but a slow pace.  They say it might be in Sept.

My posts above may have changed a bit already but most of them are still true.  What changed is after looking at mmbase (great platform too), cofax (good concept and it works, but very developer unfriendly) and opencms (very cms specific).  I think CCM is a good OSS platform for both CMS and as a platform.

I am not saying CCM is better than OpenACS all I am saying is that CCM is not as bad as most people have thought of.  I think CCM will have its uses.  I still like OpenACS but I don't have that too much negative perspective on CCM as I used to.

Collapse
Posted by David Kuczek on
Jun,

could you elaborate a little on ccm as cms? I am not sure what oacs will strategically do about cms, but I have the feeling that it might be important... Many companies simply buy a cms when they want to set up an intranet or an internet platform without thinking what exactly would fit their needs!

Is the 4.5 cms still active? Will etp be *the* cms for 4.5?

What are the most important features a cms must provide in your opinion?

Collapse
Posted by Don Baccus on
Is the 4.5 cms still active? Will etp be *the* cms for 4.5?
4.5 CMS has a ton of useful code in it coupled with incomplete integration into the [Open]ACS framework and a UI that turns everyone off. By "incomplete integration" I mean it there are a lot of rough corners in the integration (the CMS started life as an independent application and aD tried to integrate it later). Things like its own login mechanism that only works with your "screen name" which isn't required of you at the OpenACS registration end of things, the fact that the CMS code thinks it owns everything in the Content Repository. Means you can't effectively multimount it among other things, and it exposes everything from bookmarks to file storage through its UI, which is confusing.

ETP is better integrated and I think (hope?) that ETP2 will try to use more of the existing CMS code (much which is excellent) rather than roll its own.

I do think we need a more complete CMS solution than the ETP style, though ETP is a great way to put up a lot of templated information in a hurry.

Unknown to most people, I think, is that the CMS UI actually allows you to define custom content types and tie form widgets to them, allows you to attach arbitrary templates to content types or indidivual content items, etc etc. To be honest I think the CMS lets you do everything ETP does (and I know in many ways it does a lot more) ... it's just that ETP has a very straightforward and simple UI.

When 4.6 is on a roll and people are ready to talk about 4.7, we should get a subgroup underway to study CMS issues. Something you ask , "what are the most important features a CMS must provide?", would be a good planning point. Along with study of the CCM since it's now under a standard RH open source license, and maybe one or two more of the other OSS solutions Jun mentions. Not for the internals (we're in pretty good shape in that regard) but for UI, for notions of what gets exposed to various content providers, etc.

Anyway it looks like interest in CMS issues is growing again and a subgroup to explore them would be good. In a way ETP2 folk are focused on this, but they're focused more on one particular solution rather than starting from first principles, i.e. "what should a CMS look like to editors, writers etc?"

Collapse
Posted by Jun Yamog on
I don't want to get much attention on this.  That is why try to minimize the comparison of CCM and OACS publicly.

On the data model side of CMS both CCM and OACS is very similar if was not for CCM using cms_foo for its CR.  Some of the tables have been further broken down.  Like binary content and text content resides in different tables now.  As for the UI OACS needs some polishing up.  But then I don't think if we polish the UI is something that will work.

For me CMS is such a big problem space that there is no single ubber UI to satisfy.  So I think in 4.7 we break OACS CMS apart into digestable parts.  Make it more developer friendly, I doubt putting a ton on effort to the UI will move it forward.  RH is still continuing on this path using a single UI.  What OACS just needs to do is to come up with a UI that just demo's the features of CR/CMS.  And if those parts are smaller new developers can understand it and develop or extend it.  I can't even understand CR or CMS fully but now I think we really need to study the hidden treasure.  ETP was great but I just realized it redone a lot of things already present with the core.  But ETP really showed me the way on how CR works.